Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Maxdiff Simulator: First Choice Rule or Share of Preference (Logit) Rule?


We're using the Maxdiff Analyzer online tool to analyze the results of our maxdiff survey.  We're running a few simulations, narrowing down on a subset of the items used.

We're trying to determine which analysis to focus on when using the simulator (First Choice Rule or Share of Preference (Logit) Rule).  We generally understand the difference between them as explained on the Help page, but are still unsure which one is "better" or under what circumstances we'd use one versus the other.  Other than the benefit of the First Choice Rule being easier to explain to others, what are the pros and cons of each?

Thanks for your help!
asked Jul 28, 2014 by jalston (260 points)

1 Answer

+1 vote
Best answer
With MaxDiff items, the "simulator" really is just converting a set of item scores into probabilities of selection (given the context of that competitive set).

In the context of MaxDiff, I really am struggling to think of an advantage of the "first choice rule" other than it being easy to explain to others.

The logit rule (also known as Share of Preference) obtains more precision, since it leverages all the metric information for the scores within the set (rather than just focusing on which item is the one winner...irrespective of whether it wins by a lot or just by a little).
answered Jul 29, 2014 by Bryan Orme Platinum Sawtooth Software, Inc. (179,015 points)
selected Aug 1, 2014 by jalston
Got it, thank you.  We'll stick with the logit rule.